Maintaining Self Discipline
- Thread starter studentofthemarkets
- Start date
If I write the strategy myself using the strategy builder, then yes. If I write the strategy using code....which I don't know how to do, then yes.
But if I use the strategy the AI generator gives, then no, it is in a locked form that the strategy builder can't use and I have no way to change the settings.
If I learn to code and take the strategy that the AI gives and make a new strategy adding to that code, then, yes, I can change the settings.
But I'm not planning on learning to code, at least for a while.
I'm not sure if you've used the AI generator. What I've been doing is have it find a strategy using just one indicator or sometimes two or three. It will sometimes fine a really high probability strategy using those indicators you selected...but it rarely deviates from the original settings. It uses functions for entering trades that I haven't found how to do in the strategy builder..they'd have to be written in code. That is why I'm interested in saving the strategy it gives and using it on my charts....in addition to what I already do.
Right...I'm getting there...one step at a time. Thanks!
Update:
I've found that using the AI strategy generator on the 5 minute or lower timeframes hasn't been able to produce a really good strategy that stands up under longer backtesting. I think that's because my computer crashes if I try to run the generator on enough days to let it work correctly. I ended up focusing on 15 min and higher time frames. I don't think it has generated a strategy I will use by itself, but I am using some of the results of the trials to get a feel for what works and what doesn't when using a strategy that has to follow rules and yet stand the test of different market conditions.
Just for fun, I'm showing a graph of an AI generated strategy that was created from data only going back a few months. Then I ran it on the backtester from November 1, 2019 until June 3, 2021. The graph looks great at first glance, because there is so much green showing, and the end profit, for 1 ES contract was over $51,000 for about 1 1/2 years of backtesting.
However, there are a lot of drawdown areas in all of that green. From April 13, 2020 until first part of January 2021 the profits and losses oscillated around the same level. Funny thing, but looking at it, I feel like the graph would look better with a moving average on it. Maybe I've been looking at charts too long today.
If anyone wants to try to duplicate it, I'll give a hint to the indicator the strategy used: it was an oscillator...and not difficult for someone to replicate if they felt it was worth looking into.
So, although I will not be using this strategy on my charts for trading, because of the drawdowns, it was helpful to see the profit potential of a 1 hour chart. I did compare the hour chart it traded with a daily chart and found that most of those winning streaks on the graph came towards the end of a row of several strong uptrending days. So, basically, it was late in getting into the trend--this is also helpful information because I usually think it's not good to get into a trend late....I tend to want to counter trend when trends "have gone too far" yet this made quite a bit of profit when entering late in trends!
I think I'm done for a while with trying new indicators and strategies. I plan to focus on getting as far as I can with getting my trading plan as automated as possible or at least with alerts for when the best conditions are to trade it.
I've found that using the AI strategy generator on the 5 minute or lower timeframes hasn't been able to produce a really good strategy that stands up under longer backtesting. I think that's because my computer crashes if I try to run the generator on enough days to let it work correctly. I ended up focusing on 15 min and higher time frames. I don't think it has generated a strategy I will use by itself, but I am using some of the results of the trials to get a feel for what works and what doesn't when using a strategy that has to follow rules and yet stand the test of different market conditions.
Just for fun, I'm showing a graph of an AI generated strategy that was created from data only going back a few months. Then I ran it on the backtester from November 1, 2019 until June 3, 2021. The graph looks great at first glance, because there is so much green showing, and the end profit, for 1 ES contract was over $51,000 for about 1 1/2 years of backtesting.
However, there are a lot of drawdown areas in all of that green. From April 13, 2020 until first part of January 2021 the profits and losses oscillated around the same level. Funny thing, but looking at it, I feel like the graph would look better with a moving average on it. Maybe I've been looking at charts too long today.
If anyone wants to try to duplicate it, I'll give a hint to the indicator the strategy used: it was an oscillator...and not difficult for someone to replicate if they felt it was worth looking into.
So, although I will not be using this strategy on my charts for trading, because of the drawdowns, it was helpful to see the profit potential of a 1 hour chart. I did compare the hour chart it traded with a daily chart and found that most of those winning streaks on the graph came towards the end of a row of several strong uptrending days. So, basically, it was late in getting into the trend--this is also helpful information because I usually think it's not good to get into a trend late....I tend to want to counter trend when trends "have gone too far" yet this made quite a bit of profit when entering late in trends!
I think I'm done for a while with trying new indicators and strategies. I plan to focus on getting as far as I can with getting my trading plan as automated as possible or at least with alerts for when the best conditions are to trade it.
Last edited:
Just came across a Telegram thread dedicated to ONLY good news reports.
My first thoughts were that it is too off topic for my journal and should be posted in another area. But after thinking about it, I have decided that it is actually very much on topic for my journal because trading is all about making the best decisions under pressure. In order to be in a good mindset to do that, every effort should be made to live a balanced life. And that definitely includes choosing to make time to appreciate the good things in life.
Just a disclaimer: I have not gone back very far in this thread to know if there is anything offensive in it. If there is, I apologize. My understanding is that it's just good news reports from around the world.
https://t.me/theworldisgood/190
My first thoughts were that it is too off topic for my journal and should be posted in another area. But after thinking about it, I have decided that it is actually very much on topic for my journal because trading is all about making the best decisions under pressure. In order to be in a good mindset to do that, every effort should be made to live a balanced life. And that definitely includes choosing to make time to appreciate the good things in life.
Just a disclaimer: I have not gone back very far in this thread to know if there is anything offensive in it. If there is, I apologize. My understanding is that it's just good news reports from around the world.
https://t.me/theworldisgood/190
Happiness is running a backtest with a great looking green profit graph, better than the one 2 posts above.
DISSAPPOINT is running THE SAME test and getting NEGATIVE results.
Frustration is trying to pin point what could lead to differing results.
All that means more work ahead.
Apparently, when I run different Trading Hour templates on the same strategy it throws off the results...by a lot. This shouldn't happen because both Trading Hour templates were giving the indicators I was testing enough time to develop before trades began. So, either, I missed something with looking at how the indicators behaved and they were actually not giving the same readings for the two templates, or the way I'm running the test using "Standard (Fastest)" is not reliable on the 5 minute chart just using the close....which it should be reliable, from what I've read about it....or maybe there's an issue with the data it's using, either because it was run over the weekend, or because I've never downloaded any historical data. Or maybe there's another factor I'm not aware of.
Not going to work on it today, but hopefully next week I can find exactly why I'm getting differing results and if following steps to run a more accurate backtest would correct the problem.
DISSAPPOINT is running THE SAME test and getting NEGATIVE results.
Frustration is trying to pin point what could lead to differing results.
All that means more work ahead.
Apparently, when I run different Trading Hour templates on the same strategy it throws off the results...by a lot. This shouldn't happen because both Trading Hour templates were giving the indicators I was testing enough time to develop before trades began. So, either, I missed something with looking at how the indicators behaved and they were actually not giving the same readings for the two templates, or the way I'm running the test using "Standard (Fastest)" is not reliable on the 5 minute chart just using the close....which it should be reliable, from what I've read about it....or maybe there's an issue with the data it's using, either because it was run over the weekend, or because I've never downloaded any historical data. Or maybe there's another factor I'm not aware of.
Not going to work on it today, but hopefully next week I can find exactly why I'm getting differing results and if following steps to run a more accurate backtest would correct the problem.
Last edited:
Update:
The discrepancies between the backtesting results that I mentioned in the previous posts turned out to be caused by using my own Trading Hours template which begins during the Premarket. There were a couple indicators that were affected by the shortened hours. I used a shortened trading hours template so the backtest would take trades only during the time of day I am actually trading. However, having the strategy limited to trading only certain hours does the same thing and doesn't mess up the indicators.
As far as having good results and bad results...having the shortened Trading Hours template helped the short trades do better and the long trades do worse.
One key observation is that the shortened trading hours affected the number of trades taken. There were far more long trades taken and far less short trades taken when using shortened hours.
I haven't yet thought through why that would be or if I can use it to put extra conditions into the strategy to make it run better.
The discrepancies between the backtesting results that I mentioned in the previous posts turned out to be caused by using my own Trading Hours template which begins during the Premarket. There were a couple indicators that were affected by the shortened hours. I used a shortened trading hours template so the backtest would take trades only during the time of day I am actually trading. However, having the strategy limited to trading only certain hours does the same thing and doesn't mess up the indicators.
As far as having good results and bad results...having the shortened Trading Hours template helped the short trades do better and the long trades do worse.
One key observation is that the shortened trading hours affected the number of trades taken. There were far more long trades taken and far less short trades taken when using shortened hours.
I haven't yet thought through why that would be or if I can use it to put extra conditions into the strategy to make it run better.
It's been hard to get consistent backtesting results. Seems like every few months the market is changing.
Not really sure what to do about it. I've been working on building a trading robot to take trades in demo while I take the same trades live, just to keep me faithful to my plan. Now I'm wondering if I will have to readjust the robot regularly.
Not really sure what to do about it. I've been working on building a trading robot to take trades in demo while I take the same trades live, just to keep me faithful to my plan. Now I'm wondering if I will have to readjust the robot regularly.
I've decided to take this journal on a new little adventure, inspired by Saltynuts' thread https://www.elitetrader.com/et/threads/shall-we-play-a-game.356484/
and someone else who is pretty good at predicting market movement.
Trying to predict where the daily ES or NQ will end up the following day is not something I have much experience in. I will look at the daily charts to find what they are currently doing, but my focus has been on intra-day charts.
I want to see if I can learn to reliably predict the next day's daily candle. Since I am trading both the ES and NQ and can't make up my mind which one to stick with, I am going to try to predict for both of them.
I usually don't like to undertake new projects that I am likely to fail at, and this is one that I feel like I am definitely going to fail at, and I might give up after this week, depending on how badly it goes.
But I have 3 reasons for this:
1. I lack the self-discipline to do this on my own. I've tried a couple times and never lasted more than a couple days before forgetting to do it, because of all the other important things to focus on in trading.
2. I want to keep this journal going, but have been having a hard time with knowing what to post about, especially because I haven't felt much like posting what is going on in my personal trading.
3. IF I can learn to have more of a clue of what is going on in the daily chart, then I'll do better with intra-day trading, and someday maybe graduate to trading options on a swing trading basis, but that is far, far down the road.
I won't be giving out signals today for tomorrow because I'm not studying charts today. Today is my "idea" day....I like to take a break from looking at charts and that's actually when all my ideas come into my mind about trading....most of them never pan out, but today's idea is to take this journal into this new direction, and I think it's a good idea to follow through on.
So, no projecting possible prices today. I will plan to start sometime tomorrow after tomorrow's close for Tuesday. At this point, I want to try for OHLC levels. I may change that depending on how this goes.
And a warning: if anyone wants to make money....please don't take my projections into consideration when placing trades. I don't know what I'm doing, yet. If I already knew how to do it, I wouldn't need to be using this journal to learn it.
And, I will show horizontal levels for some of my reasonings, but I will also be using indicators that won't be shown, to help with the projections.
AND: If anyone wants to post here about their ideas of OHLC or anything else about the ES or NQ, that would be great too.
and someone else who is pretty good at predicting market movement.
Trying to predict where the daily ES or NQ will end up the following day is not something I have much experience in. I will look at the daily charts to find what they are currently doing, but my focus has been on intra-day charts.
I want to see if I can learn to reliably predict the next day's daily candle. Since I am trading both the ES and NQ and can't make up my mind which one to stick with, I am going to try to predict for both of them.
I usually don't like to undertake new projects that I am likely to fail at, and this is one that I feel like I am definitely going to fail at, and I might give up after this week, depending on how badly it goes.
But I have 3 reasons for this:
1. I lack the self-discipline to do this on my own. I've tried a couple times and never lasted more than a couple days before forgetting to do it, because of all the other important things to focus on in trading.
2. I want to keep this journal going, but have been having a hard time with knowing what to post about, especially because I haven't felt much like posting what is going on in my personal trading.
3. IF I can learn to have more of a clue of what is going on in the daily chart, then I'll do better with intra-day trading, and someday maybe graduate to trading options on a swing trading basis, but that is far, far down the road.
I won't be giving out signals today for tomorrow because I'm not studying charts today. Today is my "idea" day....I like to take a break from looking at charts and that's actually when all my ideas come into my mind about trading....most of them never pan out, but today's idea is to take this journal into this new direction, and I think it's a good idea to follow through on.
So, no projecting possible prices today. I will plan to start sometime tomorrow after tomorrow's close for Tuesday. At this point, I want to try for OHLC levels. I may change that depending on how this goes.
And a warning: if anyone wants to make money....please don't take my projections into consideration when placing trades. I don't know what I'm doing, yet. If I already knew how to do it, I wouldn't need to be using this journal to learn it.
And, I will show horizontal levels for some of my reasonings, but I will also be using indicators that won't be shown, to help with the projections.
AND: If anyone wants to post here about their ideas of OHLC or anything else about the ES or NQ, that would be great too.
Last edited:
Might be easier to just say looks like NQ and ES are likely to break the highs of today's Daily candles... but the NQ just did, so that's not much of a prediction. I haven't checked the ES but with a close at the daily high on such a large candle, I'd think it would be safe to say it should pop above above it at some point in the next 24 hours.
I hardly think I should be calling these predictions or projections. I guess I'll call them guesses. I won't be attempting to give a number for the New York Open.
Guesses for the NQ and ES high, low and close for tomorrow:
NQ
High 14214.00
Close: Probably somewhere close to it's high
Low: 14049.75
ES
4243.75 High
4221.25 Close
4173.00 Low
Tried to find some seasonal patterns comparing identical months for last 2 years. It doesn't seem like drops have been occurring around the same time each year. The one thing all years have in common is that they trend up. There seems to be more choppiness since the middle of February of this year to the present when comparing with February to mid-June of the previous 2 years.
These are levels that seemed to stand out on the NQ. When looking at significant drops and what happens when it breaks the previous high, shown by the yellow rectangles, there seemed to be a 1-2-3 pattern. If that plays out again and if we already made the low 2, then we are headed up or sideways to make 3.
To get the high and low, I copied the previous candle size and moved it up a little. I think the close could near the high, so I didn't make a separate estimate.
The ES has obviously mostly trended up for the last 3 years.
There doesn't seem to be the same 1-2-3 pattern after the break of the yellow rectangles on the ES. Although the daily candle was large and closed at it's high, it barely broke Friday's high. Is that a sign of weakness or just the way the ES moves?
I put the projected high just a little above Wednesday's candle from last week. I used the midpoint of today's candle for the low, partly because when looking at other days with similar price behavior it seemed the midpoint of the candle was sometimes where it would go. I went with the close as being the high from today, because when looking at other large up days coming out of a low there seemed to usually be wicks from the high of the day when it was the 2 green candle.
I hardly think I should be calling these predictions or projections. I guess I'll call them guesses. I won't be attempting to give a number for the New York Open.
Guesses for the NQ and ES high, low and close for tomorrow:
NQ
High 14214.00
Close: Probably somewhere close to it's high
Low: 14049.75
ES
4243.75 High
4221.25 Close
4173.00 Low
Tried to find some seasonal patterns comparing identical months for last 2 years. It doesn't seem like drops have been occurring around the same time each year. The one thing all years have in common is that they trend up. There seems to be more choppiness since the middle of February of this year to the present when comparing with February to mid-June of the previous 2 years.
These are levels that seemed to stand out on the NQ. When looking at significant drops and what happens when it breaks the previous high, shown by the yellow rectangles, there seemed to be a 1-2-3 pattern. If that plays out again and if we already made the low 2, then we are headed up or sideways to make 3.
To get the high and low, I copied the previous candle size and moved it up a little. I think the close could near the high, so I didn't make a separate estimate.
The ES has obviously mostly trended up for the last 3 years.
There doesn't seem to be the same 1-2-3 pattern after the break of the yellow rectangles on the ES. Although the daily candle was large and closed at it's high, it barely broke Friday's high. Is that a sign of weakness or just the way the ES moves?
I put the projected high just a little above Wednesday's candle from last week. I used the midpoint of today's candle for the low, partly because when looking at other days with similar price behavior it seemed the midpoint of the candle was sometimes where it would go. I went with the close as being the high from today, because when looking at other large up days coming out of a low there seemed to usually be wicks from the high of the day when it was the 2 green candle.